30 Years of Spinning in Circles? Use of the Paradigm Concept in Analysing Agricultural Policy Evolution

After the introduction of the paradigm concept to studies of public policy by Hall in 1993, the concept has been gratefully applied to analyse agricultural policy developments, particularly in developed states. Three paradigms are usually distinguished in the agricultural policy studies literature: the dependent/exceptionalist agriculture paradigm; the multifunctionality/ public good paradigm; and the liberal market paradigm. The concept has proved useful for describing and making sense of lacking paradigmatic policy change, although some would argue that – analysed post factum – a collection of changes taking place over a longer period of time could qualify as ‘paradigmatic’ change in agriculture (Garzon 2006) with respect to the EU Common Agricultural Policy. Overall, however, the agricultural policy domain tends to be characterized by lower-order policy changes and piecemeal reforms. It may therefore be considered remarkable that application of the paradigm concept was and remains common in this domain. But to what extent can the concept deal with analyzing lower-order changes beyond providing description or classification: can paradigms constitute causal variables? In this paper, we argue that application of the concept in explanatory analyses in the field of agricultural policy studies encounters two challenges. Firstly, there is the risk of circular reasoning: explaining the dependent variable of policy change by the independent variable of paradigm change, while measuring this paradigm change (in part) based on policy outputs – changes in the policy instrument mix. Measuring paradigm shifts based on analyzing policy discourses (problem and solution definitions and appropriate objectives of farm policy) may circumvent issues of circularity, but encounter a second problem: how can genuine discourses been distinguished from the ‘spin’ (strategic discourses) of policy actors? And hence, how can we convincingly establish that it was changes, or lack of changes, in discourse-measured ideas that causally contributed to the change, or absence of high-order change, in policy?